[Gretl-users] About the invcdf function for discrete distributions

Are we sure that the invcdf function works correctly for the case of discrete distributions? The inverse CDF method of generating a random variable using a Uniform[0,1] is valid also for discrete rv's (by using the generalized inverse), so I tried <hansl> series U = randgen(U,0,1)  # generate Uniform(0,1) series Xd = invcdf(P, 2, U)   # generate Poisson with mean 2 </hansl> but the Xd series did not have integer values, so it cannot be a Poisson rv. -- Alecos Papadopoulos PhD Athens University of Economics and Business web: alecospapadopoulos.wordpress.com/